# Central autonomic network-heart interplay in anorexia nervosa. A cross-spectral dynamic causal modeling study

## TL;DR

This study explores how brain-heart communication is altered in anorexia nervosa using advanced brain imaging techniques.

## Contribution

The study introduces cross-spectral dynamic causal modeling to investigate brain-heart connectivity in anorexia nervosa.

## Key findings

- Anorexia nervosa patients show stronger top-down signaling from frontal/insular nodes to the hypothalamus.
- Heart rate in controls correlates with bottom-up amygdala-prefrontal connectivity.
- Anorexia patients exhibit maladaptive CAN-heart dynamics through frontal-insular and brainstem circuits.

## Abstract

•CAN regulates brain–heart communication in anorexia nervosa (AN)•Effective CAN-heart connectivity in AN is investigated with cross-spectral DCM.•HR positively relates with bottom-up amygdala-prefrontal connection in controls.•HR positively associates with top-down prefrontal-amygdala causal modulation in AN.•Compared to controls, HR positively links with prefrontal-insular influence in AN.

CAN regulates brain–heart communication in anorexia nervosa (AN)

Effective CAN-heart connectivity in AN is investigated with cross-spectral DCM.

HR positively relates with bottom-up amygdala-prefrontal connection in controls.

HR positively associates with top-down prefrontal-amygdala causal modulation in AN.

Compared to controls, HR positively links with prefrontal-insular influence in AN.

The Central Autonomic Network (CAN) crucially maintains the homeostatic integrity of brain–heart communication, yet its directed interactions with cardiac control in anorexia nervosa (AN) remain poorly understood. To this end, we investigate the causal connectivity characterizing the CAN in 26 AN patients and 40 healthy controls, using a cross-spectral Dynamic Causal Modeling applied to resting state functional magnetic resonance imaging (fMRI). A Parametric Empirical Bayes framework was leveraged to estimate CAN connectivity group level differences and their linear association with group diagnosis and heart rate (HR). In patients, group connectivity differences revealed stronger top-down causal signaling from frontal/insular nodes to hypothalamus, alongside weaker connectivity from the anterior cingulate and insula to the hypothalamus and brainstem, respectively. These alterations may reflect bodily and emotional signals maladaptive processing at rest. In controls, HR was positively associated with most of the CAN connectivity, including with amygdala to prefrontal area causal influence: this might reflect a bottom-up role of amygdala in interoceptive signals processing, in absence of emotionally salient demands. In AN group, HR was negatively associated with most of the CAN connectivity, except for the connection from prefrontal to amygdala: in AN, a greater prefrontal control may emerge as a form of top-down compensatory regulation of limbic activity, at rest. Nevertheless, between-group CAN-HR differences were not found in the prefrontal-amygdala circuit. Instead, in patients, stronger top-down modulations encompassing frontal-insular and brainstem circuits resulted in driving maladaptive CAN-heart dynamics, compared to controls.